[PATCH] net: thunderx: avoid direct MTU assignment after WRITE_ONCE()

[PATCH] net: thunderx: avoid direct MTU assignment after WRITE_ONCE()

[Alok Tiwari]

The current logic in nicvf_change_mtu() writes the new MTU to
netdev->mtu using WRITE_ONCE() before verifying if the hardware
update succeeds. However on hardware update failure, it attempts
to revert to the original MTU using a direct assignment
(netdev->mtu = orig_mtu)
which violates the intended of WRITE_ONCE protection introduced in
commit 1eb2cded45b3 ("net: annotate writes on dev->mtu from
ndo_change_mtu()")

Additionally, WRITE_ONCE(netdev->mtu, new_mtu) is unnecessarily
performed even when the device is not running.

Fix this by:
  Only writing netdev->mtu after successfully updating the hardware.
  Skipping hardware update when the device is down, and setting MTU
directly.

This ensures that all writes to netdev->mtu are consistent with
WRITE_ONCE expectations and avoids unintended state corruption
on failure paths.

Signed-off-by: Alok Tiwari <alok.a.tiwari@oracle.com>
---
Note: This change is not tested due to hardware availability.
---
 drivers/net/ethernet/cavium/thunder/nicvf_main.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/cavium/thunder/nicvf_main.c b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
index aebb9fef3f6eb..ba26ba31a28bf 100644
--- a/drivers/net/ethernet/cavium/thunder/nicvf_main.c
+++ b/drivers/net/ethernet/cavium/thunder/nicvf_main.c
@@ -1589,16 +1589,18 @@ static int nicvf_change_mtu(struct net_device *netdev, int new_mtu)
 		return -EINVAL;
 	}
 
-	WRITE_ONCE(netdev->mtu, new_mtu);
 
-	if (!netif_running(netdev))
+	if (!netif_running(netdev)) {
+		WRITE_ONCE(netdev->mtu, new_mtu);
 		return 0;
+	}
 
 	if (nicvf_update_hw_max_frs(nic, new_mtu)) {
-		netdev->mtu = orig_mtu;
 		return -EINVAL;
 	}
 
+	WRITE_ONCE(netdev->mtu, new_mtu);
+
 	return 0;
 }
 
-- 
2.46.0

Re: [PATCH] net: thunderx: avoid direct MTU assignment after WRITE_ONCE()

[kernel test robot]

Hi Alok,

kernel test robot noticed the following build warnings:

[auto build test WARNING on net-next/main]
[also build test WARNING on net/main linus/master v6.16-rc3 next-20250627]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Alok-Tiwari/net-thunderx-avoid-direct-MTU-assignment-after-WRITE_ONCE/20250628-175420
base:   net-next/main
patch link:    https://lore.kernel.org/r/20250628095221.461991-1-alok.a.tiwari%40oracle.com
patch subject: [PATCH] net: thunderx: avoid direct MTU assignment after WRITE_ONCE()
config: x86_64-buildonly-randconfig-006-20250628 (https://download.01.org/0day-ci/archive/20250629/202506290114.0EPFmc8U-lkp@intel.com/config)
compiler: gcc-12 (Debian 12.2.0-14+deb12u1) 12.2.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20250629/202506290114.0EPFmc8U-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202506290114.0EPFmc8U-lkp@intel.com/

All warnings (new ones prefixed by >>):

   drivers/net/ethernet/cavium/thunder/nicvf_main.c: In function 'nicvf_change_mtu':
>> drivers/net/ethernet/cavium/thunder/nicvf_main.c:1581:13: warning: unused variable 'orig_mtu' [-Wunused-variable]
    1581 |         int orig_mtu = netdev->mtu;
         |             ^~~~~~~~


vim +/orig_mtu +1581 drivers/net/ethernet/cavium/thunder/nicvf_main.c

4863dea3fab0173 Sunil Goutham   2015-05-26  1577  
4863dea3fab0173 Sunil Goutham   2015-05-26  1578  static int nicvf_change_mtu(struct net_device *netdev, int new_mtu)
4863dea3fab0173 Sunil Goutham   2015-05-26  1579  {
4863dea3fab0173 Sunil Goutham   2015-05-26  1580  	struct nicvf *nic = netdev_priv(netdev);
f9aa9dc7d2d00e6 David S. Miller 2016-11-22 @1581  	int orig_mtu = netdev->mtu;
4863dea3fab0173 Sunil Goutham   2015-05-26  1582  
1f227d16083b2e2 Matteo Croce    2019-04-11  1583  	/* For now just support only the usual MTU sized frames,
1f227d16083b2e2 Matteo Croce    2019-04-11  1584  	 * plus some headroom for VLAN, QinQ.
1f227d16083b2e2 Matteo Croce    2019-04-11  1585  	 */
1f227d16083b2e2 Matteo Croce    2019-04-11  1586  	if (nic->xdp_prog && new_mtu > MAX_XDP_MTU) {
1f227d16083b2e2 Matteo Croce    2019-04-11  1587  		netdev_warn(netdev, "Jumbo frames not yet supported with XDP, current MTU %d.\n",
1f227d16083b2e2 Matteo Croce    2019-04-11  1588  			    netdev->mtu);
1f227d16083b2e2 Matteo Croce    2019-04-11  1589  		return -EINVAL;
1f227d16083b2e2 Matteo Croce    2019-04-11  1590  	}
1f227d16083b2e2 Matteo Croce    2019-04-11  1591  
712c3185344050c Sunil Goutham   2016-11-15  1592  
8be8a2ba18f34d5 Alok Tiwari     2025-06-28  1593  	if (!netif_running(netdev)) {
8be8a2ba18f34d5 Alok Tiwari     2025-06-28  1594  		WRITE_ONCE(netdev->mtu, new_mtu);
712c3185344050c Sunil Goutham   2016-11-15  1595  		return 0;
8be8a2ba18f34d5 Alok Tiwari     2025-06-28  1596  	}
712c3185344050c Sunil Goutham   2016-11-15  1597  
f9aa9dc7d2d00e6 David S. Miller 2016-11-22  1598  	if (nicvf_update_hw_max_frs(nic, new_mtu)) {
4863dea3fab0173 Sunil Goutham   2015-05-26  1599  		return -EINVAL;
f9aa9dc7d2d00e6 David S. Miller 2016-11-22  1600  	}
4863dea3fab0173 Sunil Goutham   2015-05-26  1601  
8be8a2ba18f34d5 Alok Tiwari     2025-06-28  1602  	WRITE_ONCE(netdev->mtu, new_mtu);
8be8a2ba18f34d5 Alok Tiwari     2025-06-28  1603  
4863dea3fab0173 Sunil Goutham   2015-05-26  1604  	return 0;
4863dea3fab0173 Sunil Goutham   2015-05-26  1605  }
4863dea3fab0173 Sunil Goutham   2015-05-26  1606  

-- 
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki